VISIO RESEARCH LABS
VRL-002 | MARCH 2026

Claims Intelligence Engine

Pre-submission validation that catches claim rejections before they cost you money. Built for South African healthcare. Powered by AI.

427
ICD-10-ZA Codes
276
NHRPL Tariff Codes
7
Scheme Profiles
13+
Validation Rules
The Problem

South African practices lose 15-20% of revenue to claim rejections

Every rejected claim means delayed payment, increased admin burden, and cash flow pressure. For a practice processing 500 claims per month, that is R60,000-R80,000 in revenue sitting in limbo — every month.

35%of rejections

Missing External Cause Codes

SA mandates V/W/X/Y codes for all injury claims. This is the #1 rejection reason — practitioners remove them to simplify coding.

25%of rejections

Insufficient Code Specificity

Using E11 instead of E11.9. Schemes require maximum specificity — 3-character codes are rejected when 4th/5th characters exist.

15%of rejections

Gender & Age Mismatches

Prostate codes on female patients. Pregnancy codes on males. Perinatal codes on adults. Often just data entry errors.

How It Works

Three steps to zero rejections

01

Upload

Upload your claims CSV or Healthbridge EDI export. Auto-detects format, maps columns, and parses data. Supports standard CSV, TSV, and Healthbridge switching format.

02

Analyze

13 core ICD-10 rules + 276 tariff codes + 7 scheme profiles + AI suggestions. Checks gender, age, specificity, external causes, unbundling, discipline, PMB eligibility, and more.

03

Fix & Learn

Auto-corrects high-confidence issues. Downloads corrected CSV. Saves results for trend tracking. Pattern learning gets smarter with every analysis.

Financial Impact

Calculate your savings

500
R800
15%
R60,000
Monthly revenue at risk
R51,000
Monthly recovery (85%)
R612,000
Annual savings
R53.9M
88 clinics / year
Validation Engine

Three layers of protection

Layer 1: ICD-10 Core

  • Code format & existence
  • Specificity requirements
  • Asterisk/dagger conventions
  • External cause codes (SA mandatory)
  • Gender & age validation
  • PMB eligibility
  • Duplicate detection

Layer 2: Tariff Cross-Validation

  • 276 NHRPL tariff codes
  • 40 unbundling rules
  • 29 discipline restrictions
  • Diagnosis-procedure matching
  • Max units per day
  • Pre-authorisation flags
  • Amount validation

Layer 3: Scheme Rules

  • Discovery Health rules
  • GEMS rules
  • Bonitas, Medshield, Momentum, Bestmed
  • Submission window (120 days)
  • CDL chronic authorization
  • Consultation limits
  • Follow-up bundling
AI Intelligence

Gets smarter with every analysis

AI Code Suggestions

Our fine-tuned medical coding models analyze flagged codes and suggest the correct ICD-10-ZA code with confidence levels and reasons. Dual-provider architecture ensures 99.9% availability.

Auto-Correction

High-confidence deterministic fixes applied automatically: non-specific codes get .9 appended, missing ECCs get X59 added, format typos (J069 → J06.9) are corrected. Medium-confidence fixes flagged for review.

Pattern Learning

Analyzes saved rejection history to identify recurring patterns. Predicts rejection probability for any code based on scheme + historical data. Gets more accurate with every analysis.

Practice Insights

Compares your rejection rate to the network average. Identifies your worst-performing schemes and most common coding errors. Generates actionable recommendations with Rand impact estimates.

Architecture

Enterprise-grade infrastructure

Dashboard Layer
Claims Analyzer | Network Dashboard | POPIA Consent Gate
API Layer (10 endpoints)
Validate | AutoCorrect | Search | Suggest | Report | History | Network | Patterns | Rules | Retention
Engine Layer (13 libraries)
ICD-10 DB | Tariff DB | NAPPI DB | Scheme Rules | Validation Engine | Auto-Correct | Pattern Learning | AI Suggest | PDF Report | CSV Export | Healthbridge Parser | Auth Guard | Types
Database + Security
ClaimsAnalysis | ClaimsRule | JWT Auth | Rate Limiting | PII Sanitization | CRON Retention
26
Files
~10,000
Lines of Code
10
API Routes
50+
Validation Rules
Security & Compliance

POPIA compliant from day one

JWT Authentication
Every API route
Middleware Protection
/api/claims/* guarded
Role-Based Access
Admin, receptionist, platform_admin
Rate Limiting
10-60 req/min per endpoint
PII Anonymization
Names → initials before storage
POPIA Consent
Checkbox required before upload
12-Month Retention
Auto-delete cron monthly
Audit Trail
userId on every analysis
VISIOCORP INNOVATION

Building intelligence for African healthcare

The Claims Intelligence Engine represents a breakthrough in South African healthcare technology. Built by Visio Research Labs, it is the first SA-built pre-submission validator with AI-powered coding assistance and scheme-specific rule intelligence.

First of its kind in SA

No existing SA tool validates claims against ICD-10-ZA specificity rules, external cause code requirements, AND scheme-specific rules in a single pass. International tools (Optum, Waystar) use ICD-10-CM — the US variant — and don't know Discovery's KeyCare GP referral rules or GEMS' 28-day chronic supply limit.

AI that understands SA coding

The AI suggestion engine is prompted with SA medical coding expertise: ICD-10-ZA conventions, NHRPL tariff codes, CMS guidelines, PMB conditions, and dagger/asterisk rules. It doesn't just suggest any code — it suggests the right SA code with scheme-specific considerations.

Pattern learning moat

Every analysis makes the system smarter. After 6 months of 88 Netcare clinics feeding real rejection data, the pattern engine will have a proprietary dataset no competitor can replicate. This is the real defensible advantage.

Healthbridge native

Auto-detects and parses Healthbridge claims switch EDI format. No CSV export needed — practices can validate directly from their existing workflow. 50+ BHF provider type codes mapped to discipline names for accurate cross-validation.

Visio Research Labs|VisioCorp|Hampton Group Associates|Netcare Primary Healthcare

Ready to eliminate claim rejections?

Access the Claims Intelligence Engine from your Netcare Health OS dashboard.